Increasing interconnections: to build or not to build, that is (one of) the question(s)
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
REPORT
Federal
Planning Bureau
Economic analyses and forecasts
Increasing interconnections:
to build or not to build,
that is (one of) the question(s)
Addendum to the cost-benefit analysis of
adequate future power policy scenarios
September 2017
Danielle Devogelaer, dd@plan.be
Avenue des Arts 47-49 – Kunstlaan 47-49
1000 Brussels
E-mail: contact@plan.be
http://www.plan.be
Federal Planning Bureau The Federal Planning Bureau (FPB) is a public agency that carries out, in support of political decision- making, forecasts and studies on economic, social-economic and environmental policy issues and ex- amines their integration into a context of sustainable development. It shares its expertise with the gov- ernment, parliament, social partners, national and international institutions. The FPB adopts an approach characterised by independence, transparency and the pursuit of the gen- eral interest. It uses high-quality data, scientific methods and empirical validation of analyses. The FPB publishes the results of its studies and, in this way, contributes to the democratic debate. The Federal Planning Bureau is EMAS-certified and was awarded the Ecodynamic enterprise label (three stars) for its environmental policy. url: http://www.plan.be e-mail: contact@plan.be With acknowledgement of the source, reproduction of all or part of the publication is authorised, except for commercial purposes. Responsible publisher: Philippe Donnay
REPORT
Federal Planning Bureau
Avenue des Arts - Kunstlaan 47-49, 1000 Brussels
phone: +32-2-5077311
fax: +32-2-5077373
e-mail: contact@plan.be
http://www.plan.be
Increasing interconnections: to build or not
to build, that is (one of) the question(s)
Addendum to the cost-benefit analysis of adequate future power policy
scenarios
September 2017
Danielle Devogelaer, dd@plan.be
Abstract – At the request of the federal Minister of Energy, this report was carried out as a follow-up
on the cost-benefit analysis published by the Federal Planning Bureau in February 2017. It constitutes
an addendum to the February study in that some additional questions impacting the Belgian production
park are scrutinized in detail. Four topics are dealt with. The first one concerns the impact of an increase
in the Belgian cross-border transfer capacity by 2 GW on the functioning of the domestic thermal flexible
park. The effect this will engender on the full load hours, the system marginal cost, CO2 emissions, the
required volumes of natural gas and employment is studied. Second, an evaluation is drafted of the cost
of keeping currently existing gas-fired power plants operational. A comparison with the cost of building
new flexible and reliable units is provided. Third, the socio-economic impact of an increased risk of a
black-out is scrutinized. The economic asymmetry this induces in relation to the costs and benefits of
conserving sufficient domestic capacity in order to comply with the legally defined Loss of Load (LOLE)
criterion of 3h is documented. Finally, the question of premature closure of currently existing Belgian
gas-fired power plants that have not come to the end of their operational lifetime yet is investigated by
means of different indicators throughout the paper.
Jel Classification – D47, L94, Q47, Q48
Keywords – Interconnections, missing money, black-out, generation adequacy
REPORT
Table of contents
Executive summary ................................................................................................ 1
Introduction ......................................................................................................... 3
1. Methodology ................................................................................................... 5
1.1. Introduction 5
1.2. Rolling horizon 5
1.3. Uncertainties 6
1.4. Electricity exchanges 6
1.5. Production technologies 6
2. Hypotheses ..................................................................................................... 8
2.1. Thermal park 8
2.2. Sun and wind 9
2.3. Electricity consumption 9
2.4. Demand response 10
2.5. International context 10
2.6. Interconnection capacity 10
3. Results ......................................................................................................... 12
3.1. Increase in the cross-border capacity 12
3.1.1. The scenarios 12
3.1.2. Coal before Gas 13
3.1.3. Gas before coal 17
3.1.4. Employment 20
3.1.5. Cost of the additional interconnections 20
3.2. Maintaining currently existing gas-fired power plants 23
3.3. Impact assessment of an increased risk of black-out 24
3.3.1. Adequate adequacy criterium 25
3.3.2. Impact of one CCGT 26
3.3.3. How many should we have? 27
3.3.4. Demand Response to the rescue 29
4. Key findings ................................................................................................... 31
5. Glossary ........................................................................................................ 32
6. References .................................................................................................... 33
REPORT
List of tables
Table 1 Combined cycle gas units in Belgium, year 2027 (MW) ····················································· 9
Table 2 Open cycle gas units in Belgium, year 2027 (MW) ·························································· 9
Table 3 Installed capacity, production (over 10 climatological years) and equivalent number of hours
at full capacity, solar PV and wind in Belgium, year 2027 ················································ 9
Table 4 Estimated costs of different interconnections to and from Belgium ····································21
Table 5 Natural gas-fired technologies, Belgium, year 2020 ······················································· 24
Table 6 Some security of supply indicators according to different scenarios, ·································· 26
Table 7 Input parameters for the DECC adequacy calculation ·····················································27
List of graphs
Graph 1 The interconnected Crystal Super Grid system, illustration ·············································· 5
Graph 2 Hourly load profile and cumulative production in Crystal Super Grid, illustration ···················· 6
Graph 3 Variability of the demand in the different test cases during one week in January, illustration ···10
Graph 4 Inframarginal rents ····························································································· 15
Graph 5 Simplified representation of load centers (dark grey), wind and solar (orange) and water (blue)
in Europe ········································································································· 22
Graph 6 Optimal level of Security of Supply ·········································································· 27REPORT
Executive summary
End of May 2017, the Federal Planning Bureau (FPB) received a formal request from the federal Minister
of Energy Mrs Marghem to elaborate further on its Cost-Benefit Analysis (CBA), a study it performed
on behalf of the DG Energy in February 2017. The goal of the CBA was to describe and quantify different
costs and benefits of varying compositions of the structural block. The structural block is a concept de-
fined by the national Transmission System Operator Elia to point to the potential investment needs in
order to construct an adequate future Belgian power system by 2027. Although different costs and ben-
efits were scrutinised amongst which the impact on the energy trade balance, on investment and system
marginal costs, emissions, etc., the Minister asked for a further unravelling of some punctual additional
topics and scenarios.
The formal request consisted in complementing the cost-benefit analysis (CBA) with, first and foremost,
an analysis of the socio-economic impact of an extension of the cross-border interconnections. Since
Belgium is an open economy with, at times, more than 25% of its current demand covered by imports,
it should be well aware of the added uncertainties linked with further increasing the amount of im-
ported electricity. The requested extension to be investigated is the 1 GW augmentation of the commer-
cial capacity of both the NEMO and the ALEGRO interconnection. A second question concerned the
cost of preserving the currently existing gas-fired power plants whilst the third question had the socio-
economic impact of an increased risk of a black-out as interest. A last question relates to the premature
closure of currently existing Belgian gas-fired power plants that have not yet come to the end of their
operational lifetime.
A number of key findings can be deducted from this analysis. First, it once again demonstrates the major
impact a fair carbon price can have on the functioning and the profitability of the Belgian gas units. If
the carbon price on the EU ETS market could trigger a switch in the merit-order between coal and nat-
ural gas, Belgian power plants will profit, run for more hours, export more electricity to the neighbours
and overall would benefit in terms of increased inframarginal rents.
However, if a rise in the carbon price does not materialize, one should try to operate on a supranational
level through credible cooperation initiatives: a reinforcement of the role of CORESO or even the Pen-
talateral Forum may in this regard be beneficiary. The point is that, although Member States have sov-
ereignty over their energy mix according to the Treaty of Lisbon, it is crucial that national phase-out
plans are coordinated across a larger ‘zone’ which is directly being impacted by the phase-out. A such
coordination may avoid black-outs, increase credibility and create an optimal level of ‘scarcity’ in the
zone.
The question whether new interconnections with both the UK and Germany will have more benefits
than costs hinges on this carbon price dilemma. In both merit-order cases (coal before gas and gas before
coal), a clear decrease in both system marginal costs (‘price’) and CO2 emissions of the Belgian electricity
sector can be observed. Nonetheless, in the former, generation by Belgian gas units is reduced and Bel-
gium becomes more and more of a hub for electricity, providing throughput or transit of power from
east to west and north to south (and vice versa). In the latter case, Belgium seems to benefit more from
1REPORT the increased interconnections, exporting vast amounts of home-made electricity towards Member States with a less advantageous, more carbon intensive electricity mix. Above that, the role of intercon- nections should not be limited to competitiveness (through price convergence) but it also spans system security of supply and integration of renewable energy sources. In terms of missing money, once again, carbon prices play a pivotal role. When carbon prices are too low, there is missing money and attracting investors becomes hazardous. When carbon prices are raised to a level that triggers the gas-before-coal switch and the legally defined nuclear phase-out is honoured, money is not missing anymore and inframarginal rents should suffice to cover a rent on top of the fixed operations and maintenance costs. Finally, if we succeed in keeping the capacity of the current operational thermal flexible park online until after the complete phase-out of all the nuclear units, generation adequacy should be assured. It is when one (or more) of the current units decides to leave the system that adequacy can no longer be guaranteed in terms of the legally defined LOLE criterium. Investments in new OCGT could mitigate the situation. Condition to trigger these investments, however, is that market design should be simul- taneously revised. Opportunities in terms of Demand Response should be carefully scrutinized because they can provide a cheaper way to guarantee security of electricity supply. 2
REPORT
Introduction
In May 2016, the national Transmission System Operator (TSO) Elia published a report on the adequacy
and flexibility needs of the Belgian power system for the horizon 2027. In June 2016, the DG Energy
organised a public consultation on the outcomes of this study. Following some of the concerns that
stakeholders that responded to the public consultation raised, the DG Energy asked the Federal Plan-
ning Bureau in October 2016 to perform a cost-benefit analysis of a selection of scenarios based on the
adequacy study of Elia.
In February 2017, the Federal Planning Bureau published its cost-benefit analysis of a number of policy
scenarios that are all consistent with an adequate Belgian power system by 2027. More specifically, five
scenarios were scrutinised, all taking into account a maximum simultaneous import capacity in Belgium
of 6500 MW by 2027. The analysis described and quantified different costs and benefits of varying com-
positions of the so-called structural block within contexts in which the merit order is either coal before
gas or gas before coal. The implications of the different capacity portfolio and import scenarios were
translated into different types of costs (investment costs, system marginal costs, consumer and producer
surplus, etc.) complemented by an impact estimation on the energy trade balance and employment.
End of May 2017, the federal Minister of Energy Mrs Marghem asked the Federal Planning Bureau to
elaborate further on this study. The formal request consisted in complementing the cost-benefit analysis
(CBA) with an analysis of the socio-economic impact of an extension of the cross-border interconnec-
tions. This additional request finds its roots in the latest Network Development Plan1 in which Elia cites
study projects aimed at reinforcing the interconnections with Germany and the UK that today are still
under construction. Since Belgium is an open economy with, at times, more than 25% of its current
demand covered by imports, it should be well aware of the added uncertainties linked with further
increasing the amount of imported electricity. The decision to further invest in additional interconnec-
tions (or reinforcements) hence should be based on a thorough cost-benefit analysis2.
This report then documents the impact of an increase in cross-border transfer capacity by 2 GW (+1 GW
to and from Germany and +1 GW to and from the UK) on the functioning of the Belgian thermal (flexible)
park. The effect on full load hours, system marginal cost, CO2 emissions, import/export patterns, re-
quired volumes of natural gas and employment is described in part 3.1.
In addition to the interconnection analysis, three other topics are being investigated. First, an evaluation
is made of the cost of maintaining currently existing gas-fired power plants operational. A comparison
with the cost of building new flexible and reliable units (OCGT’s) is provided in part 3.2.
1 Elia (2015), Federaal ontwikkelingsplan van het transmissienet 2015-2025, p. 145.
2 In the framework of the Ten-Year Network Development Plan (TYNDP), national Transmission System Operators (TSO)
and/or ENTSO-E, the European Network of TSO’s, perform CBA’s for each project included in the TYNDP. The pan-European
CBA methodology is used to assess these projects (https://www.entsoe.eu/major-projects/ten-year-network-development-
plan/CBA-Methodology/Pages/default.aspx). An example can be found at https://www.entsoe.eu/Documents/TYNDP
%20documents/TYNDP%202016/projects/P074.pdf. Above that, when a project is considered a Project of Common Interest
(PCI), information from the TYNDP is taken and the CBA is adopted as eligibility criterium. PCI’s are intended to help create
an integrated EU energy market and are considered key energy infrastructure projects by the European Commission.
3REPORT Part 3.3 then deals with the socio-economic impact of an increased risk of a black-out. The economic asymmetry this engenders in relation to the costs and benefits of conserving sufficient domestic capacity in order to comply with the legally defined Loss of Load (LOLE) criterion of 3h is scrutinized. Finally, the question of premature closure of currently existing Belgian gas-fired power plants that have not come to the end of their operational lifetime yet is looked upon. There is no chapter specifically dedicated to this theme, but throughout the paper different indicators are calculated which shine a light on the matter, the calculation of the missing money in part 3.1.2.e and 3.1.3.e being one of them. 4
REPORT
1. Methodology
The methodology used in this report is for the most part identical to (or based on) the one described in
the CBA3, part ‘Costs’. The main instrument that underlies most of the calculations is the optimal dis-
patch model Crystal Super Grid (Artelys, 2015).
1.1. Introduction
Crystal Super Grid is a unit commitment optimal dispatch model for the electricity sector that can be
used for one up to thirty three countries. It in fact minimizes total system production costs whilst align-
ing demand with supply. It contains an extensive library of both physical and financial assets (thermal
power plants, renewable energy sources, power lines, etc.) which allows a fine-grained level of detail
for analyses. The data infeed for the model mainly originates in publicly available databases like
ENTSO-E and the International Energy Agency (IEA). More specifically, the demand, the installed ca-
pacities and the thermal availabilities are obtained from ENTSO-E, the fuel costs from IEA and the de-
tailed capacity descriptions from the European TSO’s individual websites. Powerful optimization solv-
ers are used to calculate the optimal dispatch of generating facilities in the interconnected zones. Results
cover e.g. imports/exports between zones (countries or regions), marginal costs of electricity generation
and CO2 emissions.
1.2. Rolling horizon
Crystal Super Grid runs on JAVA. The computation
Graph 1 The interconnected Crystal Super Grid sys-
tem, illustration process is performed by successive optimisation
problem resolutions over a rolling horizon. This is
done to avoid perfect foresight issues at the end of
the projection period. The model computes 14-day
period (tactical horizon) problems with 7-day
steps (rolling horizon) at each iteration. This way,
each new computations’ tactical horizon overlaps
with the previous one taking into account its deci-
sions and the ensuing state of the system.
Source: Crystal Super Grid.
3 CBA is short for Cost-Benefit Analysis referring to Devogelaer, D. and D. Gusbin (2017), Cost-benefit analysis of a selection of
policy scenarios on an adequate future Belgian power system, Economic insights on different capacity portfolio and import scenarios,
Federal Planning Bureau, Report.
5REPORT
1.3. Uncertainties
Crystal Super Grid uses a (rather) deterministic approach. It does not implement the Monte Carlo method
but, through design, can operate in a similar manner. More specifically, when uncertainties come into
play, say in future demand series or intermittent production profiles, it is possible to evaluate a large
number of climatological years in one single scenario. This is done by simulating an elevated number
of test cases per scenario and by choosing its test cases strategically in such a way that ‘extremities’ are
part of the draw.
In this study, scenarios are run with ten test cases
Graph 2 Hourly load profile and cumulative produc-
tion in Crystal Super Grid, illustration each to take account of different meteorological
years and hence the influence of the weather
during a specific year on both demand and solar
and wind production. It is important to specify
that in the construction of future demand and
variable renewable production profiles, a
coherence between the two is taken into
consideration. This is ensured by including the
correlation between demand and variable
renewable production observed in different
climatological years. The model used to generate
Source: Crystal Super Grid.
the different production profiles was developed by
IAEW (Institut für Elektrische Anlagen und
Energiewirtschaft) at the university of Aachen RWTH.
1.4. Electricity exchanges
In this report, Crystal Super Grid with its hourly load profile, power plant ramp-up and emission trading
is applied to the European electricity market to study the different questions. The European electricity
market that is being considered, covers Belgium next to 32 European countries: Portugal, Spain, France,
Italy, Greece, Cyprus, Macedonia, Bulgaria, Montenegro, Serbia, Romania, Bosnia, Croatia, Slovenia,
Hungary, Slovakia, Austria, Czech Republic, Switzerland, Germany, Latvia, Luxembourg, Netherlands,
UK, Northern Ireland, Ireland, Denmark, Poland, Lithuania, Estonia, Finland, Sweden, Norway.
The commercial electricity exchanges between these different countries are modelled through intercon-
nections (NTC’s). The imported and exported volumes are calculated by the model for each scenario.
1.5. Production technologies
As regards the different power plants in this European ‘zone’, they are not modelled individually but
are aggregated into production technology categories. Two major reasons can be cited:
– There is an apparent lack of data concerning the power systems of the other 32 countries: taking into
account very detailed plant level data for the Belgian power sector but aggregating (hence, applying
6REPORT
mean efficiencies, mean production costs, etc) for the other countries would create significant biases
in the results;
– In the current version of the model, it is not possible to integrate start-up or running constraints. It
therefore does not seem to bring much added value to integrate each plant individually since the
additional time spent on integrating and running a more complex, therefore more time consuming
model will not be compensated by (much) more valuable insights.
7REPORT
2. Hypotheses
As regards the assumptions, the reader is referred to the CBA published in February. Nonetheless, for
the sake of easy comprehension some crucial assumptions are repeated, next to the description of hy-
potheses that are specific to this study.
2.1. Thermal park
Since this study takes the legal context as a given, notably the law on the progressive nuclear phase-
out4, the nuclear power plants are not part of the exercise.
Concerning the remaining thermal park, two types of production can be distinguished: fatal (or must-
run) and dispatchable. As regards the fatal production, two categories are integrated in the model: bio-
mass and cogeneration. The production plan of these technologies is not optimised by the model be-
cause they are considered to be units that have to produce, no matter how low (or high) the market
price. The installed capacity of cogeneration equals 2938 MW with an annual average electricity pro-
duction of 17 TWh. The installed biomass capacity is not expected to change compared to the situation
of today: it is set at 1281 MW with an annual production around 8 TWh.
Upon decision of the Cabinet of the federal Minister of Energy, the dispatchable thermal park taken into
consideration is made up of the existing gas-fired power plants that are currently in the market. This
means that the generation units belonging to the strategic reserve are excluded from the study. The
operational Belgian thermal production park as of 2017 is therefore integrated in the scenarios covering
horizon 20275. The source used is the database constructed by Elia in the framework of the determina-
tion of the need for strategic reserves for the winter 2017-2018.
One exception is made for Drogenbos: according to a communication with ENGIE, the Drogenbos
OCGT (46.8 MW) is being decommissioned, whilst the CCGT (460 MW) is converted to an OCGT com-
prising of 230 MW nameplate capacity made up of 2 GT’s of 115 MW each. Hence, Drogenbos is with-
drawn from the list of CCGT and adapted in the list of OCGT. Table 1 summarizes the list of Belgian
combined cycle gas units, Table 2 the open cycle gas units.
4 January 31, 2003, Loi sur la sortie progressive de l'énergie nucléaire à des fins de production industrielle d'électricité, http://www.ejus-
tice.just.fgov.be/cgi_loi/loi_a.pl?language=fr&caller=list&cn=2003013138&la=f&from-
tab=loi&sql=dt=%27loi%27&tri=dd+as+rank&rech=1&numero=1
5 As the Cabinet specifically demanded to take this park into consideration, the question whether this park will still be available
by 2027 given its age or how many investments will be necessary to keep it operational is not dealt with in this publication.
8REPORT
Table 1 Combined cycle gas units in Belgium, year 2027 (MW)
Production unit name Type Production capacity (MW)
AMERCOEUR 1 GT CCGT-GT 270
AMERCOEUR 1 ST CCGT-ST 150
ANGLEUR TG 31 CCGT-GT 39
ANGLEUR TG 32 CCGT-GT 39
ANGLEUR TV33 CCGT-ST 39
HERDERSBRUG GT1 CCGT-GT 159
HERDERSBRUG GT2 CCGT-GT 159
HERDERSBRUG ST CCGT-ST 162
INESCO GT1 CCGT-GT 48.5
INESCO GT2 CCGT-GT 44.8
INESCO ST CCGT-ST 44.8
KNIPPEGROEN STEG CCGT 305
MARCINELLE ENERGIE TGV CCGT 405
RINGVAART STEG CCGT 357
SAINT-GHISLAIN STEG CCGT 350
T-POWER CCGT 425
ZANDVLIER POWER CCGT 384
TOTAL 3381
Source: Elia (2016).
Table 2 Open cycle gas units in Belgium, year 2027 (MW)
Production unit name Type Production capacity (MW)
ANGLEUR TG 41 GT 63
ANGLEUR TG 42 GT 63
DROGENBOS GT GT 230
HAM 31 GT 56
HAM 32 GT 56
IZEGEM GT 22
TOTAL 490
Source: Elia (2016), ENGIE.
2.2. Sun and wind
As regards the installed capacity of variable renewable energy sources, this is identical to what is im-
plemented in the Base case (Elia, 2016a). Table 3 summarizes some indicators used in this report.
Table 3 Installed capacity, production (over 10 climatological years) and equivalent number of hours at full capacity,
solar PV and wind in Belgium, year 2027
Installed capacity (MW) Mean annual production (GWh) Equivalent full-load hours (h)
Solar PV 4988 5563 1115
Wind* 5854 15852 2708
Source: Crystal Super Grid.
Note: * both on and offshore wind.
2.3. Electricity consumption
In the model, the demand is represented by an hourly load profile. In order to guarantee the supply-
demand equilibrium of the system, the electricity demand has to be satisfied at all moments in time by
an equal power production (being generated somewhere in the system). The challenge to meet this bal-
ance is linked to the variability of the residual demand.
9REPORT
The uncertainty on the future power demand is
Graph 3 Variability of the demand in the different
test cases during one week in January, illus- taken into account by simulating different test
tration
MW cases that are correlated with the production pro-
files of wind and sun as to attempt to accurately
capture the variability of the residual demand (see
also part 1.3). Graph 3 provides an illustration of
the fluctuating demand in the different test cases
implemented per scenario.
The average annual growth rate of the electricity
demand between 2014 and 2027 equals 0,6% (Elia,
2016b). In the year 2027, the average annual de-
mand (over the 10 test cases) equals 92,9 TWh.
Source: Crystal Super Grid.
2.4. Demand response
Following some remarks from stakeholders after the publication of the CBA, it was decided to further
finetune this option as to more accurately capture the potential of this valuable source of flexibility.
Whereas in the CBA, it was assumed that 1096 MW of Demand Response (DR) would be available by
2027 in coherence with the adequacy study from Elia (Elia, 2016a), the potential was revised upwards
and adapted to the number used in the Addendum of Elia (Elia, 2016b). Next to that, the price was
revised downwards. This study hence accounts for 1600 MW of DR6 with an activation price set at 175
€/MWh for 50% of this potential, 275 €/MWh for the other 50%.
2.5. International context
It is important to recall the international setting of this study. At the specific request of the Cabinet, the
international framework in which the analyses are conducted is taken identical to the one described in
the CBA of February. This means that, for neighbouring countries, the same capacity portfolio (hence,
phase-out calendar) is implemented as is described in the CBA (Devogelaer and Gusbin, 2017).
2.6. Interconnection capacity
The commercial exchanges between countries are modelled through nodes in which every country is
represented by one node. The exchanges are limited by the net transfer capacities (NTC). Losses are not
considered in the model but a small part (0.1%) is deducted with any energy exchange as to avoid sym-
metries and equivalent solutions. The maximum simultaneous import capacity for Belgium in 2027 is
foreseen to have been increased up to 6500 MW.
6 This potential can be seen as an extrapolation of the total MR volume with a CAGR of 5% (E-Cube, 2017).
10REPORT
The first part of the study then integrates a further extension of the interconnections. This means that
an additional 2000 MW is considered which corresponds to the extension of both the NEMO and the
ALEGRO link with 1 extra GW7, constituting a maximum simultaneous import capacity of 8500 MW in
2027. This option is already mentioned in the latest Federal Network Development Plan of Elia (Elia,
2015) and was investigated in the adequacy study as a sensitivity (Elia, 2016a).
7 Although an investment of 2 GW in physical capacity does not necessarily lead to an increase of 2 GW in commercial capacity,
we subscribed to the analysis published in part 4.3.2 and 5.4.3 (Elia, 2016a) where it is stated that increases in the commercial
capacity are taken into account and the maximum simultaneous import capacity is lifted from 6500 MW to 8500 MW.
11REPORT
3. Results
3.1. Increase in the cross-border capacity
The first research question concerns the impact of an increase in the cross-border transmission capacity
by 2 GW (+1 GW to and from Germany and +1 GW to and from the UK) on the functioning of the Belgian
thermal flexible park. The effect this will induce on the full load hours, CO2 emissions, system marginal
cost, the required volumes of natural gas and on employment will be documented in this part. Through-
out this chapter and in a transversal manner, the question whether this option may lead to a premature
closure of currently existing Belgian gas-fired power plants that have not come to the end of their oper-
ational lifetime yet is being investigated by means of different indicators, the calculation of the missing
money being one of them.
3.1.1. The scenarios
To tackle this question, four scenarios are constructed. They are broadly inspired on the contexts de-
scribed in the CBA (Devogelaer and Gusbin, 2017).
The first scenario, called Gas2017, is a scenario in which the Belgian production park is identical to the
Base_Gas scenario, although with two exceptions: more DR is available in the system8 and there is no
structural block anymore. In fact, it is more accurate to state that the structural block which, in the CBA,
was filled with different types of technologies (either gas-fired or decentral), is replaced by the gas-fired
generation units that are currently part of the market in Belgium in 2017. Basically, this choice is made
because in this report, the subject is not approached from the angle of security of supply or generation
adequacy, but from an economic point of view in which the profitability (and survival) of current units
is investigated. More specifically, potential changes in the profitability level by adding interconnection
capacities is looked upon. To summarize: 3381 MW CCGT (see Table 1) and 490 MW OCGT (see Table
2) are integrated in the model9.
The second scenario called Gas2017_HighCO2 is identical to the first but with one difference: in the merit
order, gas comes before coal. To put it differently: the CO2 price is (significantly) raised10. The analysis
of this scenario can deliver insightful results because if the future carbon price is sufficiently high, it
becomes interesting to switch from coal-based to gas-fired power plants, the latter seeing their number
of running hours increase (dramatically). (Partly) replacing coal and lignite will have an effect that is
not contained to Belgium: it can cross borders in terms of induced changes in production levels, import
and export patterns and CO2 emissions all over Europe.
The third and fourth scenario, called Addinterco and Addinterco_HighCO2 respectively, are identical to
scenario one and two but with an extension/reinforcement of the interconnections between Belgium and
8 See also part 2.4.
9 This apparently lower level of gas-fired units compared to the structural block of 4.4 GW in the CBA does not jeopardize the
legal criterium of generation adequacy because a simultaneous increase in the level of DR is considered (Artelys, 2017).
10 In the CBA, this can be interpreted as the Clima_Gas scenario but in which the structural block is being deleted and the current
gas-fired park is added.
12REPORT
the UK and Belgium and Germany. In other words, by 2027, NEMO II and ALEGRO II are installed and
operational and the maximum simultaneous import capacity is being lifted to 8500 MW.
Before plunging into the results of these scenarios, it is important to note that Elia in its adequacy study
(Elia, 2016a) already performed a sensitivity analysis on the import capacity. In part 5.4.3 of its study,
the results of an analysis in which the maximum simultaneous import capacity of Belgium is increased
to 8.5 GW are described. According to their findings, the necessary volume of the structural block in
2027 can decrease by 1.5 GW when 2 GW of import capacity is added to the Belgian system. It is none-
theless important to stress that Elia worked with steps of 500 MW and that this result is valid only in
their Base scenario. More specifically, interconnections indeed bring a benefit in reducing the adequacy
risk for as long as there is sufficient available production capacity on CWE/European scale. This benefit
dilutes if the latter is not the case. For the Base case as we use in our study, this could mean that, for
adequacy reasons, the structural block (which can be interpreted as the gas-fired park we will concen-
trate on in this analysis) may be smaller. Hence, the park of 2017 will not be needed entirely from a
security of supply point of view.
In the next sections, scenarios are compared two by two, meaning that the scenario Gas2017 is compared
to the scenario Addinterco and the scenario Gas2017_HighCO2 is compared to Addinterco_HighCO2. Re-
sults are reported below.
3.1.2. Coal before Gas
a. Full load hours
In the first analysis, the situation of constructing additional interconnections is looked upon in a context
in which the CO2 price hovers around 17 €/tCO2. When we place the two scenarios next to one another
(Gas2017 vs Addinterco), we notice that, in the case of 2 GW additional interconnections, 0.8 TWh less
electricity is being produced by Belgian CCGT. Belgian CCGT therefore will count 232 less full load
hours (FLH), tumbling from 4739 to 4507 FLH (-5%). Belgian OCGT will lose out on 64 full load hours
going from 787 to 723 FLH (-8%).
This electricity is being replaced by more imports (+3 TWh). At the same time, we notice an increase in
exports (+2.2 TWh). In a situation with 8.5 GW of maximum simultaneous interconnection capacity, we
import more from Germany and the UK (apparently the 1 GW interconnection was a bottleneck), but
less from the Netherlands and France. We also tend to export more to all our neighbouring countries,
especially to the UK. The export to the UK grows from 1.9 TWh in a situation with NEMO I to 4 TWh
with NEMO II operational. This can largely be attributed to the time difference between the two coun-
tries. We also note that Belgium becomes more and more of a hub for electricity, providing throughput
or transit of power (from Germany to the UK for example).
b. Required volumes of NG
Less production by CCGT and OCGT means less natural gas required to supply these units. The supply
of natural gas destined to the power sector is reduced by 1.7 TWh-GCV, going from 33.2 TWh in Gas2017
to 31.5 TWh-GCV in Addinterco. This volume represents 2.4% (resp. 3.5%) of gas used for generating
13REPORT
power in 2010 (resp. 2015). Compared to the projected natural gas imports in 2027 in a reference scenario
(FPB, 2017, forthcoming), it is less than 1%.
c. CO2 emissions
Of course, differences in production and import/export patterns translate in differences in CO2 emis-
sions11. For Belgium, because less natural gas is being burned to produce power, the CO2 emissions
decrease by 0.3 Mt. Overall, on a European scale, they increase by 0.8 Mt. This increase is due to a rise
in the production of coal- and lignite-based power in Europe: coal production increases by 1.7 TWh,
lignite by 0.1 TWh. The CO2 emissions released by burning coal are only for a small part compensated
by a lower production of CCGT (-1.3 TWh) and OCGT (-0.5 TWh). Countries that use more coal are
principally Germany and Poland, less gas is used in a.o. Belgium and the UK.
d. System MC
The system marginal cost can be seen as a proxy for the “price” in an energy-only market design, but
some nuances have to be taken into account:
– The system marginal cost is calculated as an average over the course of a year (8760h) and over the
ten test cases: it therefore is not a momentary shot of the EPEX screen;
– It relates to the price of the commodity, not to the bill the end consumer receives at the end of the
billing period;
– It depends on a number of assumptions as to the hypotheses relative to the price of the fuel, the
conversion efficiency (which is taken to be 56% for existing CCGT, 59% for newly built CCGT and
34% for OCGT), the variable O&M costs and the price of CO2. The marginal production cost of an
existing CCGT (resp. OCGT) then amounts to appr. 60 €/MWh (resp. 105 €/MWh) in 2027.
Comparing the two situations, we notice that the construction of the two additional interconnections
will have a declining effect on the average system marginal cost. Installing a NEMO and ALEGRO ex-
tension will trigger a downward pressure on system marginal costs of 0.33 €/MWh on average. Alt-
hough this impact may seem rather modest, it has to be reminded that it is the result of the averaging
out over the 10 test cases and the 8760 hours. Zooming in on test case level, hourly price differences up
to 160 €/MWh can be detected, meaning that in the Addinterco scenario the price for 1 MWh of electricity
at certain moments in time may be 160 € lower than in Gas2017. This happens for example when the
latter has to rely on the more expensive DR option (priced at 275 €/MWh) whilst in the former, a foreign
OCGT located in the UK can supply power at a lower price.
In absolute terms, average system marginal costs will decrease from 71.1 €/MWh in Gas2017 to 70.7
€/MWh in Addinterco. This will reflect on the consumer surplus: the consumer surplus in Belgium will
increase by 39 M€, or 39 M€ less has to be spent by Belgian consumers on the purchase of the commodity
electricity (excluding taxes, tariffs, etc.) in 2027. At the same time, production surplus will decrease be-
cause production of gas-fired units diminishes and the average price the producers receive for gener-
11 It is important to recall that emissions from cogeneration are not reported in the model.
14REPORT
ated electricity will be lower: this leads to a loss of 21 M€ in production surplus. For the UK and Ger-
many, the opposite is happening. Their marginal costs increase, hence diminishing their consumer sur-
plus, but because of increased production and higher wholesale prices, their production surplus over-
compensates the loss in consumer surplus.
e. Missing Money
Missing money is a term that is used to describe a situation in which prices during critical hours are too
low, leading to net revenues (revenues net of fuel and other operating costs) being inadequate to sup-
port the efficient level and mix of generating capacity. This will engender underinvestment in generat-
ing capacity, too many hours when capacity is fully utilized, too much reliance on non-price rationing,
and too high a probability of a network collapse (Cramton and Stoft, 2006; Joskow, 2007).
To get a grasp of the missing money, several methods are applied. A first method is inspired by Elia
(2016a). In part 5.5 of the adequacy study, an analysis of the economic parameters of gas-fired units is
provided. It goes about calculating the cumulative distribution of the inframarginal rents of CCGT. It is
important to recall that the Elia study (as this analysis) only covers revenues earned in the day-ahead
market. Other revenues are not considered although they could represent a more stable stream of in-
come for CCGT.
In this part of the study, we calculate the in-
Graph 4 Inframarginal rents
framarginal rents for the existing collection of
CCGT in the Gas2017 scenario12. According to the
different test cases, different point estimates are ob-
tained. They seem to vary widely: the average in-
framarginal rent by 2027 reaches 81 M€ for 1 GW
installed, but the minimum is at 36 M€. In the latter
case, it becomes difficult to just hang on to the unit
since Fixed Operations & Maintenance (FOM) are
estimated to lie in the interval [12.5-36] (see below,
part 3.2). With an inframarginal rent of 81 M€, FOM
can be covered, but this amount does not neces-
Source: DG Trésor France.
sarily incite new investments. The decision to in-
vest crucially depends on the WACC. According to Elia (2016a), the sum of the FOM and the annuity
with a WACC of 4% should reach 75 M€ for 1 GW (or 75 €/kW), a WACC of 10% necessitates 115 M€
(or 115 €/kW).
Another method to calculate the missing money consists in comparing the system marginal cost with
the levelized cost of electricity (LCOE) for natural gas-fired power plants. The LCOE in fact is the net
present value of the unit cost of electricity over the lifetime of a generating asset. It is often taken as a
proxy for the average price that the generating asset must receive in a market to break even over its
lifetime. This LCOE is not a uniform number: it differs across technologies and even within technologies,
12 We did a similar exercise for the Addinterco scenario. As the average ‘price’ in the latter is slightly lower, conclusions remain
valid.
15REPORT
variations can be noted. The LCOE depends on a number of parameters amongst which the paid interest
rate, fuel cost, load factor and the technical lifetime.
Two sources are used to determine the LCOE of Belgian CCGT. The first is the joint publication of the
International Energy Agency and the Nuclear Energy Agency (2015). The LCOE for CCGT in Belgium
in the year 2020 is estimated to be 114 $/MWh (103 $/MWh) when its load factor attains 50% (85%)13.
When we compare this LCOE with the average “price” we obtain through our model (71 €/MWh) with
a calculated load factor of 54%, we see that the LCOE cannot be repaid through revenues from the day-
ahead market alone.
Another source is Albrecht et al. (2014). In this publication, an LCOE for CCGT is taken between 74 and
150 €/MWh. At the lower end of the interval, the load factor is assumed to be rather high. In that case,
the CCGT seems to be marginally viable, but only slightly attractive for investors: they should just be
able to recoup their investment. At the higher end, no investor will dare to jump.
It is important to point out that, if there is missing money, some firm thermal generators could be en-
ticed to quit the market because of unavailability of FOM recovery. This may happen in a rather ‘disor-
derly’ manner. When it materializes, supply shrinks at identical demand which leads to prices experi-
encing an upwards pressure. This could even result in sustained periods of above long run marginal
cost pricing, hence impacting in a positive manner the revenues of those generators who decided to stay.
In terms of inframarginal rents earned by the remaining CCGT, we calculated that the demise of one
CCGT could increase the rents by [3-15]%. In part 3.3, we analyse the impact this could have on security
of electricity supply.
A point to stress is that our calculations only concern the year 2027. If future years hold more promise,
it might be possible to reel in investors. But since wholesale power prices are very low today and that
they are predicted to remain low for the years to come (Elia, 2016a), this should mean that the years
after 2027 should be especially profitable with very high wholesale power prices to make up for the
below-average years. This is rather unlikely since 1) politicians will take measures to avoid/mitigate the
high price impact; 2) this would incite more people to self-produce and store their own electricity (be-
come a prosumer), hence decreasing residual demand and future power prices.
Also, whether gas-fired power plants will have more rocky times during the intermediate ten-year pe-
riod and whether they will survive up until 2027 is not the object of this study14. By 2027, the nuclear
phase-out, as legally defined, will have created room for these natural gas units to operate. If the nuclear
phase-out, for whatever reason, is postponed, this will impact the calculations. More specifically, the
inframarginal rents and profitability prospects for gas-fired units will decrease because of the merit-
order effect of prolonged nuclear units (Devogelaer and Laine, 2016).
The second analysis described in part 3.1.3 zooms in on a future in which the carbon price is significantly
higher and triggers a gas before coal switch in the merit order.
13 Note that both fuel and carbon costs are estimated to be higher in the IEA/NEA publication. By comparing the results with
other countries having lower fuel and/or carbon costs, our conclusion seems robust.
14 Elia (2016a) gives some insights.
16REPORT
3.1.3. Gas before coal
a. Full load hours
In this second analysis, the situation of constructing additional interconnections is looked upon in a
context in which the CO2 price hovers around 55 €/tCO2. Compared to the previous situation, net elec-
tricity production levels in Belgium are significantly higher: when we compare the scenarios Gas2017
and Gas2017_HighCO2, almost 10 TWh additional power will be generated by Belgian CCGT in the case
of higher carbon prices. Throughout Europe, coal and lignite power plants lose ground in higher CO2
priced scenarios because of the imminent penalization proportional to their carbon content. The gener-
ation in these carbon intensive plants will be replaced by an increase in the production of CCGT. Coun-
tries that have natural gas in their capacity portfolio will augment its production, Belgium being
amongst them. Therefore, Belgian gas-generated power will increase, raising CCGT load factors, and
will flow to Member States that burn coal like the Netherlands and Germany.
In the case of the 2GW additional interconnection (scenario Addinterco_HighCO2), the amount of elec-
tricity produced by Belgian CCGT even increases somewhat further (+37 GWh) but OCGT generate
slightly less (-31 GWh). In terms of functioning, Belgian CCGT will have 11 more full load hours, going
from 7656 to 7667 FLH (+0.1%). OCGT, on the other hand, will lose out on 63 full load hours, going from
784 to 721 FLH (-8%).
Import and export patterns change. When NEMO II and ALEGRO II are in place, exports from Belgium
towards all our interconnected neighbours increase to attain 15.5 TWh compared to 11 TWh in
Gas2017_HighCO2. Exports to Germany even double: they grow from 4.1 TWh in Gas2017_HighCO2 to
8.7 TWh in Addinterco_HighCO2. Imports towards Belgium also thrive. Where they reached 31 TWh in
Gas2017_HighCO2, they now are at 35.5 TWh with imports from the UK to Belgium almost doubling:
they grow from 7.3 TWh in Gas2017_HighCO2 to 14.3 TWh in Addinterco_HighCO2. What we can learn
from these additional flows is that, in the case of a reinforcement of the interconnections with both
Germany and the UK, Belgium will import more from gas producing countries (UK), less from neigh-
bours having a more polluting mix (e.g. Netherlands, Germany) and will export more to all its neigh-
bours, specifically to Germany.
b. Required volumes of NG
When a higher CO2 price is assumed, additional interconnections have no effect on natural gas supply
to the power generation sector. Although the higher production of the CCGT exceeds the lower produc-
tion of the OCGT, the difference in conversion efficiency (lower for OCGT than for CCGT) results in
similar (absolute) changes in gas volumes. In other words, the additional gas used in the CCGT is ex-
actly compensated by the lower gas consumption of the OCGT. In both scenarios, natural gas supply to
the power sector is estimated to be 53 TWh-GCV in 2027.
17REPORT c. CO2 emissions As regards CO2 emissions15, it is first instructive to cite the additional amount of CO2 produced in the Belgian electricity sector by going from a context with relatively low carbon prices to one in which the price of CO2 triggers a shift in the merit order. In the latter, because of the dramatic increase in CCGT activity, CO2 emissions in Belgium expand by 3.2 Mt. It should however be brought to the attention that the power sector is part of the European Emission Trading System (EU ETS) which is a cap-and-trade system. The emission reduction objective therefore is determined at European level by placing a cap on emissions whilst no target for ETS sectors is defined at Member State level. In practice, this means that when emissions in a certain country increase, additional permits have to be bought by the producers in that country on the destined auctions. This does not necessarily mean that the EU ETS cap will be sur- passed, rather the price of the permits will, in a normally functioning market, rise making the emitting of carbon (hence the electricity price) more expensive. When we then compare the situation with and without NEMO II and ALEGRO II, we see that in the former, CO2 emissions are marginally lower: in the scenario Addinterco_HighCO2, they shrink by 5.7 kt (or 0.006 Mt). However, on a European scale, the effect is amplified. Because of the shift in the MOC triggered by the higher CO2 price, adding additional interconnections to and from Belgium will give room to a further replacement of the most polluting fuels and technologies. All over Europe, 1.9 TWh of coal, 1.9 TWh of lignite and 0.5 TWh of OCGT are being replaced by an additional 4.2 TWh of CCGT, giving way to a decrease in CO2 emissions. This effect is calculated to represent a decline of 2.5 Mt CO2 on European scale. d. System MC Comparing the two scenarios, we notice that the construction of the two additional interconnections will have a declining effect on the average system marginal cost. Installing a NEMO and ALEGRO ex- tension will trigger a decline in system marginal costs of 0.31 €/MWh on average. It is important to specify that this average is taken over all hours of the year (8760h) and over all ten test cases. Zooming in on test case level, hourly price differences up to 129 €/MWh can be seen, meaning that at certain hours, the presence of NEMO II and ALEGRO II will allow Belgian wholesale power prices to be 129 €/MWh lower than without those extensions. In absolute terms, average system marginal costs will decrease from 97.1 €/MWh to 96.8 €/MWh. This will reflect on the consumer surplus. The consumer surplus in Belgium will increase by 33 M€, or 33 M€ less has to be spent by Belgian consumers on the purchase of the commodity electricity (excluding taxes, tariffs, etc.) in 2027. At the same time, production surplus will decrease because the average price the producers receive for generated electricity is lowered. The production surplus will be 20 M€ lower. The UK sees its production shoot by 4.4 additional TWh. Although its system MC is a bit higher (thereby lowering consumer surplus), the production surplus will be raised substantially. Germany, on the other hand, sees its production shrink by 2.6 TWh and its system MC decrease. This leads to a drop in the German production surplus triggered by both a price and volume effect. The consumer surplus, on the other hand, increases considerably due to lower wholesale power prices. 15 It is important to recall that emissions from cogeneration are not calculated in the model. 18
You can also read
